EVALUATION OF THREE TYPES
OF FISH REARING PONDS
By ROGER E. BURROWS, Fishery Research Biologist
Fish and Wildlife Service
and HARRY H. CHENOWETH,
Hydraulic Engineer
University of Washington
RESEARCH REPORT 39
Fish and Wildlife Service, John L. Farley,
United States Department of the Interior, Douglas McKay,
Director
Secretary
UNITED STATES GOVERNMENT PRINTING OFFICE : 1955
For sale by the Superintendent of Documents, United States Government Printing Office
Washington 25, D. C. - Price 1 5 cents
CONTENTS
Page
Introduction
.......... 1
Determination of Hydraulic Conditions, by Harry H
.......... 2
Chenoweth
Factors influencing model studies
.......... 2
Determination of model size
.......... 3
Description of prototypes
.......... 4
Foster-Lucas pond
.......... 4
Circular pond
.......... 4
Raceway pond
.......... 5
Hydraulic characteristics of the ponds
.......... 7
.......... 7
Flow patterns in models and in prototypes
Flow pattern in the Foster-Lucas pond ......
8
Flow pattern in the circular pond
.......... 8
Flow pattern in the raceway pond
........ 11
Comparison of flow patterns
........ 12
Short circuiting in models and in prototypes
........ 14
Short circuiting in the Foster-Lucas pond
........ 17
Short circuiting in the circular pond
........ 17
........ 17
Short circuiting in the raceway pond
Conclusions
........ 18
Correlation of Hydraulic Conditions with Physical and
Biological Characteristics, by Roger E. Burrows
........ 19
Factors affecting efficient pond operation
........ 19
Carrying capacity
........ 19
Disease inhibition
........ 23
Food distribution
........ 24
Cleaning efficiency
........ 25
Comparative pond efficiencies
........ 26
Summary and conclusions
........ 27
Literature cited
........ 29
II
EVALUATION OF THREE TYPES
OF FISH REARING PONDS
INTRODUCTION
Artificial propagation subjects
salmon and trout to abnormal conditions. Confinement of large numbers of fingerlings in relatively
small volumes of water may create
an environment definitely detrimental to optimum growth and
development. The several types of
fish-rearing ponds now in use differ
both in the manner of maintaining the fish and in efficiency of
operation.
A study of the hydraulic, biological, and physical characteristics of
several types of ponds was undertaken with the ultimate objective of
either improving present types or
developing new designs. This
paper is concerned only with the
first phase of the investigation,
namely, the development of methods of evaluating fish ponds based
on their hydraulic characteristics.
The paper will show that biological
and physical conditions in ponds
are dependent on the hydraulic
characteristics, and further, that the
hydraulic characteristics can be
predicted by the use of models.
Three types of ponds were selected for test : the Foster-Lucas,
the circular, and the raceway.
These particular types were chosen,
not because they were considered
either the best or the worst, but because more actual operating information and more experimental data
have been accumulated for them
than for some of the other types.
They also are more or less typical of
the fish-rearing ponds in general
use. Each of these types has been
constructed in various sizes and
with considerable variation in the
basic design. On the assumption
that trial and error would have
eliminated impractical alterations,
the ponds tested were selected from
large groups recently constructed.
Small rearing ponds on the Grand
Coulee project supplied the FosterLucas type of rearing pond used
during these investigations and the
Little White Salmon Station
(Wash.), the raceway—both constructed by the U. S. Fish and
Wildlife Service. The Marion
Fork Station of the Oregon State
Fish Commission provided the circular type of pond. Thus, although the findings of this study
with regard to each type of rearing
pond apply only to the specific modification of the type studied during
this investigation, minor alterations
in design and operation do not alter
the basic characteristics sufficiently
to invalidate general comparisons.
1
In the hydraulic studies, the flow
patterns in both ponds and models
were determined by the use of floats
and dyes. Comparisons of the
degree of mixing and short circuiting were ascertained by the injection of dye into the influent and the
measurement of the time of appearance and the concentration of the
dye in the effluent. The biological
and physical characteristics of the
three types of ponds were determined from experiments conducted
at the Salmon-Cultural Laboratory,
Entiat, Wash., and from evaluations made by other investigators.
Model studies were incorporated
in these investigations to determine
the model size necessary to reproduce the hydraulic characteristics
of fish-rearing ponds, and to facilitate various portions of these and
future studies. The use of scale
models will eliminate the necessity
for the construction of full-sized
ponds to evaluate new pond designs
and the alterations necessary for
the improvement of established
types.
The design and scale drawings of
the models were prepared by Scott
Bair. Trevor Watson constructed
the pond models. Dr. Robert
Rucker and Charles Wagner assisted in the hydraulic and chemical
evaluations.
Results of this investigation are
presented in two parts : The first is
concerned with determination of
the hydraulic conditions which
exist in the three types of ponds
studied and the degree to which
these conditions can be duplicated
in models ; the second correlates
hydraulic conditions with known
biological and physical characteristics of the ponds.
DETERMINATION OF HYDRAULIC CONDITIONS
By HARRY H. CHENOWETH
FACTORS INFLUENCING MODEL STUDIES
From their previous experience
with ponds and similar pools of
water and knowledge of other investigators' findings, it was logical
for the present writers to assume
that criteria could be established
for evaluating the biological and
physical performance of fish-rearing ponds from their hydraulic
characteristics. Therefore, they
simultaneously sought such criteria
and developed model techniques
that could be used to determine the
hydraulic characteristics of either
2
altered ponds or new designs. A
study of the flow pattern in each
type of pond was required for both
the correlation of the hydraulic and
biological characteristics and the
determination of the degree to
which a model would reproduce the
hydraulic characteristics.
Many problems concerning the
flow of fluids are too complicated
to be solved by mathematical analysis alone, and scale models were
developed to solve these problems.
The structure which the model rep-
resents is called the prototype.
The model should be geometrically
similar to the prototype (in certain
cases a distortion in the vertical
direction may be desirable), but
this geometrical similarity is not
enough to ensure that the fluid motion of the prototype will be accurately reproduced in the model. If
the direction of flow and the relative velocities are the same at corresponding points in both the model
and its prototype, the flow is said to
be kinematically similar. The
force required to change the direction of a moving streamline depends on the mass as well as the
velocities of the particles composing the fluid. If densities and velocities are proportional, the model
is said to be dynamically similar to
the prototype. The dynamic forces
are often the predominant forces,
but sometimes viscous drag and surface tension are not negligible.
Complete similitude requires that
all of the properties of the fluid in
the model be related correctly to the
corresponding properties of the
fluid in the prototype. The proper
density, viscosity, et cetera, of the
fluid in the model depends on the
geometric scale ratio between model
and prototype, and on the density,
viscosity, et cetera, of the fluid in
the prototype. It is apparent that
with the limited number of fluids
available for use in the model
complete similarity is practically
impossible. Fortunately, absolute
similitude is not necessary to obtain practical results. In any
particular hydraulic problem, one
law is usually the dominating one
and other effects may be ignored
if they are small, or the results of
following the major law can be adjusted to take care of the secondary
influences.
DETERMINATION OF MODEL SIZE
The flow pattern in fish-rearing
ponds is governed principally by
inertial forces and, hence, models
of these ponds should follow
Froude's law (essentially, the velocities should be reduced by the
square root of the scale ratio) . If
Froude's law is to be followed, the
velocity in the model should be to
the velocity in the prototype as the
square root of the linear dimensions
of the model is to the square root of
the linear dimensions of the prototype. For example, if the model is
one-ninth the size of the prototype,
the velocities in the model should
be the square root of one-ninth, or
one-third the velocities in the prototype. The velocities in a 1: 10
scale model should be adjusted so
that they are 0.316 times those in
the prototype. The flow, being the
product of velocity and area, would
be proportional to the 2.5 power of
the linear-dimension ratio. The
flow through a 1: 10 scale model
should therefore be 0.00316 times
that through the prototype.
The largest of the secondary influences on fluid motion is viscous
drag and, unfortunately, this influence is not negligible. If viscosity
affects the flow pattern in the prototype, any model smaller than the
3
prototype will have a distorted flow
pattern. If the flow in the prototype is fully turbulent (i. e., all
velocities high enough that viscosity is not a factor), there is a
critical model size above which no
appreciable distortion occurs and
below which a gradually increasing
distortion of the flow pattern results as the model size is decreased.
In certain regions of all fish
ponds, velocities are low enough
that viscosity affects the flow pattern. Because of these viscous effects, the flow is not fully defined by
geometry throughout the prototype, and hence there is no critical
model size. The smaller the model,
the greater is the effect of viscous
drag. A scale ratio of 1: 10 was
selected as the most practicable for
these model studies although it was
known in advance that some distortion of the flow pattern would occur
due to viscosity. Models which
were exact replicas of the FosterLucas, circular, and raceway prototypes were constructed one-tenth
normal size.
DESCRIPTION OF PROTOTYPES
The prototypes selected for study
were of modern construction and
incorporated the best features of
each of the pond types. Great variation existed in shape, size, and
water inflow between the three
types of ponds.
FOSTER - LUCAS POND
The Foster-Lucas pond w a s
roughly oval-shaped, and had a
center partition wall. The inside
measurements of the prototype
tested were 76 by 17 feet. The
pond was operated with a mean
depth of 3 feet, and its capacity
was 28,000 gallons. During the
tests, the flow was adjusted to 202
gallons per minute, which approximated normal operating conditions. The water circulated around
the center partition wall, and
finally passed out through screens
located on each side of the
partition wall near the center of
the pond. In the particular Foster4
Lucas pond studied, water entered
through perforated influent pipes
running transverse to the major
axis of the tank. Water was introduced through eighteen %6-inch
orifice holes drilled in each of the
two 4-inch standard-weight influent headers. The headers were rotated so that the axes of the jets
were depressed 45° below the horizontal. Figure 1 illustrates the
model of the Foster-Lucas pond
used in this study.
CIRCULAR POND
The circular pond tested had a
25-foot inside diameter and a flat
bottom. The depth of water was
311
/2 inches, making the capacity
of the pond 9,630 gallons. Water
was admitted through a single nozzle that delivered 105 gallons per
minute, and was directed tangentially to the peripheral wall but
tilted downward at an angle of 35°
to the horizontal. The outlet
Figure 1.—Model of the Foster-Lucas pond showing the dye-iniection tank and mercury differential
gage. The gage is connected across a diaphragm orifice, and is used to set flow to the desired
quantity.
screens were located at the center
of the pond. The circular pond
model is shown in figure 2.
RACEWAY POND
The raceway pond tested was 8
feet wide, 80 feet long, and averaged 2 feet deep. The capacity of
this pond was 9,650 gallons. The
bottom had a very slight slope to-
ward the outlet end to facilitate
drainage when the pond was being
cleaned. Water was admitted to the
pond through slots between boards
along the side of an elevated flume
header. Each side of the pond was
fed by a separate slot. The total
inflow to the pond was 490 gallons a
minute. Figure 3 shows the raceway pond model.
5
Figure 2.—Model of the circular pond.
Figure 3.—Model of the raceway pond. The two plastic plates in the foreground were added
to facilitate capture of the effluent.
6
HYDRAULIC CHARACTERISTICS OF THE PONDS
FLOW PATTERNS IN MODELS AND
IN PROTOTYPES
When the models were operated
at the same Froude number as the
prototypes, the same general patterns of flow were observed. The
greatest discrepancies were in those
regions where velocities were low.
These regions were more sluggish
in the models than in the prototypes. This behavior is explained
by the fact that in low-velocity regions of the models inertial forces
were relatively small, and hence
viscosity of the fluid readily dampened out eddies and even hindered
their formation. Dye injected in
these regions of the model was
pulled into nonturbulent streamlines, or streaks. Although this
typical phenomenon of viscosity was also observed in the same
regions of the prototype, the tendency was not so pronounced. Lowvelocity regions in the prototype
will be relatively lower in the
model, and conversely, when predicting the flow characteristics of
a proposed prototype from observations made on a model, one should
expect a little less difference in relative velocities in the prototype
than in the model.
Before proceeding to a detailed
discussion of the flow pattern in
each type of pond, a few remarks
dealing with flow characteristics in
open conduits or channels are in
order. At very low velocities, the
flow is nonturbulent. The movement of the fluid down the conduit
may be thought of as being laminar,
or a sliding of layer on layer, al334317 0-55
2
though this concept is somewhat of
an oversimplification. The velocity
profile is parabolic. Velocities at
the fixed boundary are zero, but increase rapidly with the distance
from the boundary until the maximum is reached a short distance
under the free surface, and then decrease slightly because of surface
tension.
At higher mean velocities, the
flow becomes turbulent except for
a thin layer near the bottom (see
fig. 4). The velocity profile is somewhat the same as in nonturbulent
flow except that the velocities near
the fixed boundary are relatively
larger. The velocities at the fixed
boundary are still zero, however.
If the velocity in the conduit is
gradually increased, this laminar
layer next to the fixed boundary
will become thinner, and if the velocity becomes great enough, the
laminar layer will disappear altogether. The flow is then said to
be fully turbulent.
Since in fully turbulent flow the
velocity against the fixed boundary
is a relatively large percentage of
the mean velocity in the conduit,
wall roughness becomes an important factor. . If fully turbulent
FREE SURFACE
"N-- SURFACE
TENSION EFFECT
BOUNDARY LAYER
Figure 4.—Velocity distribution in partially
turbulent flow.
7
flow is to be expected, then the
roughness of the wall of the model
should be a scale reproduction of
t h e prototype's roughness. It
should be pointed out that the three
modes of flow (degrees of turbulence) just described are typical of
straight conduits. Curvature, obstructions, sudden changes in section, et cetera, will alter the flow
pattern locally. These local phenomena are likely to be a major
item in shaping the flow pattern of
most fish-rearing ponds.
Flow pattern in the Foster-Lucas
pond.—In the particular FosterLucas type pond used in these
studies, the momentum imparted to
the water by the 36 influent jets
caused the water to circulate around
the pond with a maximum velocity
of about 0.8 feet per second. The
general flow pattern is shown in
figure 5, a. The flow pattern near
the ends of the ponds was unstable
since observations over a period of
time revealed continuously changing details. The high velocities
near mid-depth in the reach just
upstream from the turn carried
through to the wall, causing a roll
off the end wall. Streaks from the
dye crystals placed at the bottom,
as shown at the right in figure 5, b,
were straight across the pond rather
than on a curved path paralleling
the curved end wall. After making
the bend at the end of the pool,
the main flow was near the outside
wall. The bulk of this outer stream
was carried on around the pond, but
some of the flow found its way to
the outlet screen and some of it
turned into a large eddy behind the
partition wall. A spoonful of dye
8
crystals was scattered in this area
and the resulting coloration clearly
defined the extent of the eddy (see
fig. 5, b). The similarity of flow in
the model and prototype observed
near the end of the center partition
wall is shown in figure 6.
The study of the flow pattern in
the Foster-Lucas pond revealed
three undesirable conditions: A
large eddy behind the partition
wall, short circuiting, and a roll at
each end of the pond. The eddy
was primarily objectionable because
of the low velocities in the area.
The mixing action in the eddy
contributed to short circuiting,
although the rolling of the water
at the ends of the pond and the introduction of the influent by a series
of orifices extending completely
across the width of the pond probably played a more important role
in the short-circuiting action.
Flow pattern in the circular
pond.—Water admitted at the
periphery of the circular pond
through the tangentially placed
nozzle should flow spirally around
the pond gradually approaching the
screen and outlet located at the center of the pond. The actual flow
pattern in the circular pond under
study was quite different. The outer
zone, near the wall, was a region
of high velocity and intense mixing.
Inside this outer zone, the main
flow was circular with a secondary
spiral motion superimposed (see
fig. 7, d). While the main flow revolved around the vertical axis of
the tank, a slow twisting motion
carried the water near the bottom
inwardly toward the screen. Some
of the water carried in by this sec-
I
0"
RADIUS
SCREENS AND OUTLET PIPE
INFLUENT ORIFICES
...111,•■■
A) GENERAL FLOW PATTERN IN FOSTER-LUCAS POND
B)
SOME FLOW PATTERN DETAILS
Figure S.—Flow pattern in Foster-Lucas pond. a. General flow pattern; b. Some flow pattern
details.
DYE LIES AGAINST END OF
WALL AND FEEDS OFF THE
3" STEEL CHANNEL
UPSTREAM EDGE
Ng
VORTICES
MODEL
VORTICES
PROTOTYPE
Figure 6.—Comparison of flow patterns in model and prototype near the end of the central
partition wall of the Foster-Lucas pond.
ondary spiral passed through the
screen and thence to the outlet. The
greater portion, however, rose as it
revolved around the screen (see
fig. 7, b).
While the inward flow along the
bottom contributed to short circuiting, in this particular case short circuiting may be advantageous. Set-
tling matter is picked up by this inward current and is carried to the
screen. The objectional hydraulic
features of this circular pond were
the large peripheral mixing zone
and the torus-shaped (doughnutshaped) dead region. The circular
axis of this torus lay roughly midway between the bottom of the pond
9
C)
NOZZLE
7
UPWARD
SPIRAL
B) FLOW PATTERN
AROUND SCREEN
A) PLAN VIEW OF CIRCULAR POND
VERY LOW VELOCITY
I DEAD AREA
-41
D)
.
1
SLOPE
4'
3
8 STD. PIPE OUTLET
MEDIUM VELOCITY
SECTION
Figure 7.—Flow pattern in circular pond. a. Plan view; b. Flow pattern around screen; c. Nozzle;
d. Section.
and the free surface and about a
third of the radius out from the
center of the pond, as shown in figure 8. This region is described as
dead because of the lack of interchange of water between it and the
rest of the pond, rather than a lack
of actual motion. This entire region revolved around the central
axis of the pond at a reasonably
high velocity, though this velocity
was less than that near the pond
wall or near the screen (see fig. 7).
Since this was a relatively sluggish
area in the prototype, this tendency
toward low velocity was accentu10
ated in the model. Figure 9 shows
the model a short time after the
flow of dye had been stopped.
Clear water had moved from the
outer mixing zone to the screen by
passing under the torus-shaped
area. Much of the dye that finally
found its way into the dead area
was held there and the area shows
dark on the photograph.
The driving force of the tangential jet must be transmitted over a
relatively long distance by shear
stresses in the fluid to reach this
region. Making the pond smaller
in diameter or deeper should tend to
Figure 8.—Model of the circular pond photographed shortly after dye had been added to the
influent. Note the dead area that did not color and the typical viscous streaks near the central
............ n.
alleviate the sluggishness in this
region. Because of the retarded
boundary layer, the velocity distribution is less uniform in shallow
ponds than in deep ones. Since a
moving mass will travel in a
straight line rather than a curved
path unless acted upon by some external force, the high velocity above
mid-depth tends to travel tangentially to the curved path and thus
tends to move to the outside wall.
Because the momentum of the water
near the top is greater than that
near the bottom, this top water is
deflected downward at the wall and
then moves in against the lowvelocity water at the bottom. This
action explains the secondary roll
observed in the model and the
prototype.
Flow pattern in the raceway
pond.—The only available raceway
pond was one of a battery of 20
ponds. Unfortunately, the flow
pattern of the pond studied was not
typical of that in the other ponds.
It was estimated that one slot was
delivering about twice the flow of
the other (see fig. 10, d). This lack
of symmetry of the influent stream
caused an unsymmetrical flow pattern in the pond (see fig. 10, a).
The unsymmetrical intake was duplicated in the model and a similar
flow pattern was observed. The
1
influent slots were about 3 /2 feet
above the water surface of the pond
11
Figure 9.—Photograph of circular pond taken 45 minutes after the dye-injection period. Note
that the dead area shown in figure 8 as clear now contains an appreciable amount of dye.
(see fig. 10, c), and so the falling
sheet of water struck the surface
with a velocity of about 15 feet per
second. The turbulence created was
still evident slightly below the midlength of the pond. The bottom
velocities were directed upstream.
The influences of viscosity were
clearly noticeable in the region just
upstream from the screen in the
prototype, and for a relatively
greater distance in the model. It
was thought that a more distinctive
flow pattern could be obtained for
model-verification purposes if water
were admitted from one slot only.
This was done and the flow pattern
is diagrammed in figure 10, b.
Since the raceway pond does not
have the advantage of recirculation
as do the Foster-Lucas and circular
12
ponds, its mean velocity will be low
unless an extremely large quantity
of water is available. High inlet
velocities at the upper end of the
raceway type of pond contribute to
short circuiting, but they can easily
be eliminated by the use of a suitably baffled intake. Such a baffle
added to the intake of the model resulted in a definite improvement.
COMPARISON OF FLOW PATTERNS
The location of a dead region
may be more important than the
extent. A raceway pond fed
through a well-baffled intake will
have dead areas only at the fixed
boundaries. If the quantity of
water passing through the pond is
large, this boundary layer will be
thin. If the flow through the race-
AU
20
C
A)
/el
FLOW PATTERN IN A RACEWAY POND
THIS FLOW PATTERN WAS OBTAINED WITH THE UNSYMMETRICAL INTAKE SHOWN IN
DEAD
AREA
BOTTOM
—
D)
\
/
\
B)
FLOW PATTERN WITH HALF OF THE INTAKE BLOCKED
NARROW SLOT
THIN SHEET
WIDE SLOT
HEAVY)
FLOW
3
D) DETAIL OF
UNSYMMETRICAL INTAKE
R)
LONGITUDINAL. SECTION
Figure 10.—Flow pattern in raceway pond. a. Flow pattern in raceway pond with unsymmetrical
Intake; b. Flow pattern with half of intake blocked off; c. Longitudinal section of pond; d. Detail of unsymmetrical intake.
way is small, the boundary layer
may be of consequential thickness
and, unfortunately, would extend
the full length of the pond.
The doughnut-shaped dead area
in the circular pond lies above a
layer of water that is moving toward the screen with a medium
velocity. Settling solids that drop
out of the dead region may be
picked up by the undercurrent and
carried toward the screen. This
desirable feature is somewhat offset
by the accessibility of this dead
area to all fish and by its remoteness
from the outlet screen.
The dead regions behind the center partition wall in the FosterLucas pond are about as accessible
and as remote from the screens as
in the circular pond, but the region
extends from the top to the bottom
of the pond. Furthermore, the
dead area is fed by water coming
from the direction of the outlet
screen, and water that has passed
through the dead area is picked up
by the main flow and is carried
around to the intake on the opposite
side of the wall (see fig. 5, b). The
mixing zones in the raceway and
circular ponds are confined primarily to the inlet region, whereas
in the Foster-Lucas pond there is
considerable mixing at the intake,
the turn at each end of the pond,
and at each end of the large eddies
behind the center wall.
13
SHORT CIRCUITING IN MODELS
AND IN PROTOTYPES
If efficient use is to be made of
the water in a fish-rearing pond, the
influent water should pass through
all parts of the pond before reaching the outlet. If water passes
quickly from the inlet to the outlet
without circulating to all parts of
the pond, short circuiting is said
to take place. Short circuiting is
exhibited by all ponds to some degree and is due to differences in
velocities and lengths of the stream
paths, and it is accentuated by high
inlet velocities and by mixing due
to eddies. Short-circuit studies are
usually made on model ponds operated in accordance with Froude's
law. A quantity of dye or salt is
injected into the influent and the
concentration of the added substance in the effluent is observed at
the end of various time intervals
until virtually all the substance
has passed from the pond. From
the data thus obtained a curve of
concentration against time may be
plotted. If these curves are plotted
in dimensionless form (see fig. 11),
they are a means of comparing the
hydraulic characteristics of ponds
of different shapes.
The abscissa for a point on the
dimensionless curve is the relative
time, i. e., the actual time divided
by the theoretical detention time.
The ordinate is the relative concentration computed by dividing the
observed concentration by the concentration that would be obtained
if the dye or salt injected at the
intake were instantaneously dispersed throughout the pond volume.
This type of plot is independent of
14
the amount of dye or salt used and
is also independent of the detention
time of the pond.
The ideal pond would have a
relative-concentration, relative-time
curve as shown in figure 12. Any
actual curve will depart considerably from this ideal shape, but the
less its departure the better the
pond from a hydraulic standpoint.
If duplicate runs made under the
same conditions yield different
curves, the flow pattern of the pond
is unstable. The more unstable the
flow pattern, the more unpredictable are the hydraulic properties of
the pond. Instability will lead to
erratic performance, and is therefore undesirable. Low velocities
in a pond are conducive to instability. If there are dead spaces in
which the liquid plays little or no
part in the general displacement
through the pond, the effective pond
volume will be less than the true
volume. Consequently, the effective detention time will be less than
the theoretical detention time, and
the relative time to the centroid of
the area under the relative-time,
relative-concentration curve will be
less than unity. Since the relative
time to the centroid is unity except
for ponds with dead spaces, the relative time to the centroid is an indicator of the amount of dead space in
the ponds. If there is any interchange between the dead spaces and
the main flow, there will be a long
"tail" on the end of the curve. The
relative time to the center of the
area under the curve is usually less
than unity and decreases the greater
the short circuiting. The actual
time to the center of the area is the
4
PROTOTYPES
CIRCULAR
1.0
s?
-
FOSTER- LUCAS
RACEWAY, TEST I
RACEWAY, TEST IC
0.5
1.0
1.5
2.0
2.5
EFFLUENT
CONCENTRATION
UNIFORMLY
CONCENTRATION
RATIO OF ACTUAL TIME TO THEORETICAL DETENTION TIME
MODELS
0.5
1.0
1.5
2.0
RATIO OF ACTUAL TIME TO THEORETICAL DETENTION TIME
Figure 11.—Relative time-concentration curves for the prototype (above) and models (below).
probable flowing-through time,
since half the particles of fluid
move through the pond in less time
than this and half pass through in
more time than this.
Attention is called to the fact that
the terms "centroid" and "center"
are not synonymous. For convenience, centroid may be thought
of as the center of gravity, whereas
center means that there is as much
area on one side of a given point as
there is on the other. A line
through the center divides the area
into two equal parts, while a line
through the centroid may or may
not divide the area into equal parts.
The centroidal axis is a balancing
15
4
WIDTH EQUAL TO THE
RELATIVE TIME OF INJECTION
0.5
10
RELATIVE TIME
Figure 12.—Dye-concentration-time curve of
an
to the influent nozzles in the prototypes may explain some of the variation between the model curves and
the corresponding prototype curves.
Since the observed time started with
the beginning of the dye-injection
period, one-half the injection time
was subtracted from the probable
flowing-through time and from the
effective detention time shown in
table 1.
ideal fish-rearing pond.
Table
axis and its position depends on the
distribution of the area as well as
its magnitude. The relative time
of the initial appearance of the dye
or salt is also a measure of short
circuiting.
Tests both of models and of prototypes were made in which gentian-violet dye was injected in the
influent by means of a small pressure tank, a converted weed burner
(see fig. 1). The concentration at
the outlet was determined with a
photocolorimeter. Theoretically,
dye-injection time should be infinitesimal. Actually this is not
possible, nor for that matter, desirable. All ponds are at least slightly
unstable, and a finite dye-injection
time gives a somewhat average picture, which is desirable unless instability is being studied. The injection period for the models was
30 seconds; for the prototypes it
varied from 5 to 9 minutes. The
large amount of dye necessary in
the tests of the prototypes made a
shorter period impractical. The
fact that the relative injection times
in the models and in the prototypes
were slightly different and that the
dye was injected somewhat closer
16
1.—Relative flowing-through time in
the
three types of ponds.
[Time for ideal pond taken as 1]
Foster- CircuLucas
lar
Time of first appearance of
dye at outlet
Effective detention time__ _
Probable flowing-through
time
Raceway
0.023
. 83
0.055
.88
0.345
1. 11
.57
.58
.92
The ideal pond would have a relative time of unity for each item
shown in table 1, and the smaller
the actual value the worse the hydraulic performance from t h e
standpoint of short circuiting.
Stability is also an important hydraulic characteristic, but statistical data on stability are expensive
to obtain. Some instability was
observed in all of the ponds when
the flow pattern was traced. As
previously mentioned, theoretical
considerations indicate that instability will be greatest ill low-velocity ponds, and so one would expect the greatest instability in the
raceway type of rearing pond.
The values shown in the table are
the averages of several tests and
although the deviations from the
average values were greatest in the
raceway pond, all the relative times
from the individual tests were
larger than those obtained for the
circular pond or the Foster-Lucas
pond. The comparisons made in
the table, therefore, seem valid. A
relative effective detention time
greater than unity for the raceway
pond may be explained partly by
the instability of this pond due to
its low velocity.
Comparison of each of the model
curves with the corresponding prototype curve shows that invariably
the maximum concentration of dye
reaches the outlet in a relatively
shorter time in the models than in
the prototypes, and that the relative concentration of dye is higher
in the models. This condition
means that the relative size of the
dead spaces is larger and that the
short circuiting is worse in the
models, as is to be expected, because
of the relatively greater influence
of viscosity in the case of the
models. This condition agrees
with the results found in tracing
flow patterns. An important point
to note in predicting the characteristics of a proposed prototype from
the type of tests. under discussion is
that the predicted curve should be
sketched slightly to the right and
with ordinates of about 90 percent
of those for a 1: 10 scale model.
Short circuiting in the FosterLucas pond.—Direct observation of
the flow pattern in the Foster-Lucas
pond revealed serious short circuiting. Studies in the dye-concentration-time relationship gave a quantitative evaluation of this characteristic. Both the relative time of the
first appearance of the dye at the
outlet and the relative effective detention time are measures of the
short circuiting. As shown in the
table, both of these criteria mark the
Foster-Lucas pond as hydraulically
inferior to either the circular or the
raceway pond. The relative probable flowing-through time is a measure of the extent of the dead areas.
The extent of these regions was
about the same in the Foster-Lucas
and circular ponds, but the fact that
the relative effective detention time
was greater for the Foster-Lucas
pond indicated more interchange
between these dead regions and the
regions of main flow. This increased interchange is further substantiated by the long tail on the
Foster-Lucas curves.
Short circuiting in the circular
pond.—Studies in the dye-concentration-time relationship confirmed
the existence of a large dead area
in the circular pond (the observed
doughnut-shaped region) and, furthermore, showed this region to occupy about the same percentage of
the pond volume as did the dead
areas in the Foster-Lucas pond.
While the short circuiting was considerably more in the circular pond
than in the raceway pond, it was less
than that in the Foster-Lucas pond
by both criteria shown in the table.
It should be remembered that much
of the short circuiting in this pond
occurred along the bottom and was
directed to the outlet. This assisted
in making the circular pond selfcleaning and, therefore, was not
entirely undesirable.
Short circuiting in the raceway
pond.—Considerable difficulty was
encountered in injecting the dye
17
evenly across the slots of this pond.
The other two prototypes and all
of the models were fed by pipes into
which dye could be injected easily.
The influent slots of the raceway
pond were fed by a large open channel, but during the dye-injection
period of the first test on the prototype most of the dye entered the
pond through one slot. Therefore,
a second test was run. Most of the
dye entered through the other slot
at the beginning of this run but was
fairly evenly distributed during the
latter part of the injection period.
This difficulty was not experienced
when testing the models, and both
the prototype and model tests indicate that the dead regions in this
type of pond are smaller than in
the other two types tested. The
dead areas are confined principally
to the quite stable boundary layers.
Dye trapped in these regions is held
for some time resulting in a fairly
long tail on the concentration-time
curves (see fig. 11). Despite the
unfavorable intake characteristics
of the raceway pond tested, the
superiority of this type of pond in
regard to short circuiting and dead
areas is clearly shown by the curves.
CONCLUSIONS
The hydraulic characteristics of
fish-rearing ponds may be studied
satisfactorily by 1: 10 scale models.
The data obtained from the model
should be corrected for the minor
influence of viscosity when predicting the performance of the
prototype.
The three specific types of ponds
studied during the investigation
have serious hydraulic defects of
such a nature that only major
changes would correct them. Were
the intake of the raceway pond
modified by baffling, the low initial
cost of this type of pond might justify its use where a large quantity
of water is available. Rarely will
velocities be high enough to make
the raceway pond self-cleaning.
Since the natural path of a mov-
18
ing mass is a straight line, and since
the flow against any fixed boundary
is retarded, it is difficult to see how
any circular pond having recirculation can be freed of short circuiting
and mixing. The curved ends of
the Foster-Lucas pond gave characteristics similar to the circular pond
and the eddies behind the partition
wall were even more serious.
It is, of course, easy to find fault
with existing ponds, but much
harder to offer improvements. The
authors felt that before they could
undertake the task of searching for
improvements, they should determine and study the faults of existing ponds and develop criteria that
could be used in the development of
better fish-rearing ponds.
CORRELATION OF HYDRAULIC CONDITIONS WITH
PHYSICAL AND BIOLOGICAL CHARACTERISTICS
By ROGER E. BURROWS
The actual operating character- sions and assumptions have been
istics of established types of fish- made and criteria established by
rearing ponds have become known which rearing-pond types have
over the years. Most of the reports been evaluated. In this section of
on pond operations in the literature, the report, the hydraulic conditions
however, are not accompanied by determined for the three types of
comparative data from which ac- ponds used in this study—the
curate evaluations of the ponds can Foster-Lucas, the circular, and the
be made. As a result, comparisons raceway—will be correlated with
of operating characteristics of var- the known biological and physical
ious types of ponds under compa- characteristics of each pond, and
rable conditions are practically un- criteria established by which alteraobtainable. Despite the dearth of tions in pond design may be
comparative data, certain conclu- evaluated.
FACTORS AFFECTING EFFICIENT POND OPERATION
Carrying capacity, disease-inhibiting qualities, food-distribution
characteristics, and cleaning efficiency, have been selected as the
major criteria to be used in evaluating the efficiencies of the three types
of ponds. The carrying capacity
of a pond includes both the inflow
measured in gallons per minute and
the concentration expressed as
pounds of fish per cubic foot of
water. Disease inhibition is evaluated on the basis of the differential
in resistance of the fish to diseases
endemic to the water supply, which
under normal operating conditions require routine prophylaxis
for control. The food-distribution
characteristic is defined as the ability of a pond to distribute the food
throughout its area. Cleaning efficiency is determined by the disposition of excrement and debris in re-
lation to the outflow and the effort
required to remove this material
from the pool. Although other
factors also influence pond efficiency it is believed that they are
closely associated with one or more
of these four major criteria.
CARRYING CAPACITY
In a comparison of the carrying
capacity of the raceway and circular types of pools, Cobb and Titcomb (1930), Surber (1936), Prevost (1941), and Davis (1946), all
state that the circular pond is
superior to the raceway both with
regard to the inflow of water required and the pounds of fish that
may be carried per cubic foot of
water. Experiments conducted by
this laboratory cast some doubt on
their conclusions. If the gallonsper-minute inflow is used as the
19
criterion, then the circular pond can
carry more fish per cubic foot of
water, but if the inflow is ignored
the carrying capacity of the raceway appears to be superior to that
of the circular pool.
The hydraulic conditions that exist in a raceway are very similar to
those in a deep trough, as indicated
by model studies. Experiments
conducted in deep troughs with
blueback salmon as the test animals
indicated that the carrying capacity of this type of trough was in
excess of 5 pounds per cubic foot
for this species. In these tests, the
water inflow was increased as the
load increased. Actually, physical
factors such as screen and drain
capacities were overtaxed by the
increased water inflow required to
meet the oxygen demand and
forced abandonment of the experiment before a reduction in growth
rate was attained. The poundage
per trough in the heaviest stocking
was in excess of 100 pounds (more
than 5 pounds of salmon per cubic
foot) when the experiment was
abandoned. Davis (1946) found
that rainbow trout had a normal
growth rate in shallow troughs at
poundages in excess of 5 pounds per
cubic foot.
Experiments with chinook salmon to determine carrying capacities of circular ponds indicated
that the capacity of the circular
pond studied was slightly in excess
of 1 pound of fish per cubic foot of
water. Contrary to popular belief,
carrying capacity did not increase
with larger fish but was purely a
function of weight and volume.
Pond capacities were determined
20
for fish ranging from 280 to 40 per
pound.
Species differences may invalidate this comparison between raceways and circular pools, although
the work of Johnson and Gastineau
(1952) and Palmer et al. (1952) indicate that the growth of chinook
salmon is not inhibited in circular
tanks. It is not anticipated that
raceways could carry loads of 5
pounds per cubic foot unless the
normal input of water was very
greatly increased. It may be expected, however, that loads in excess of 1 pound of fish per cubic foot
could be exceeded providing the
inflow, screen area, and drainage
facilities were adequate. Rodgers
(1949) mentions stocking a particular type of raceway at the rate of
1.6 pounds of fish per cubic foot, but
he made no comparisons between
this raceway and other types of
ponds.
No actual comparisons of the
Foster-Lucas pond with either the
raceway or the circular pool have
been possible. Another factor—
that of disease—has prevented the
carrying of fish loads of more than
one-third pound per cubic foot in
Foster-Lucas ponds available for
test on the Grand Coulee project.
It is doubtful, however, if this type
of pond would approach the carrying capacity of a circular pool under
even the most favorable conditions.
Elements which influence the
carrying capacity of a pond are
believed to be the available oxygen,
carbon-dioxide content, and accumulation of metabolic waste products. Oxygen content of the water
is affected by the amount of inflow-
ing water, the path of water flow identical water capacities, but were
through the pond, and possibly the not comparable with respect to the
distribution of the fish in the pond. amount of water introduced. Five
The amount of oxygen available to times the amount of water rethe fish is limited by the oxygen quired by the circular pool was used
contained in the water supply and in the raceway under actual operatthe amount of water introduced, ing conditions. On the basis of
plus the surface area of the pond. water path alone, under the condiEllis et al. (1946) state that 'an tions of these tests, the raceway was
oxygen content of 5 p. p. ni. is ade- superior to the circular pond and
quate for fish life and that of 3 the circular pond was superior to
p. p. m. is lethal. Davis (1946) the Foster-Lucas pond.
Davis (1946) states that, because
points out that the oxygen requirement of a fish varies with its of the more even distribution of fish
activity and is correlated with feed- in the circular pool, they have a
ing. If the water inflow is limited better opportunity to extract oxyor the oxygen content reduced, there gen from the water and that this
is danger of an oxygen depletion factor is responsible for the greater
approaching the lethal level that carrying capacity of the circular
may well be a limiting factor in the pool. This may very well be true,
but it should not be considered
carrying capacity of a pond.
The path of water flow also may in terms of the available oxygen
affect the amount of oxygen avail- but rather in terms of the inhibiting
able to the fish. In the hydraulic effect of oxygen absorption prodeterminations, the path of water duced by accumulations of metawas measured by the amount of bolic waste products. Irvin et al.
short circuiting that occurred. If (1941) and Black and Black (1950)
the inflowing water is not com- have demonstrated that increases in
pletely available to all the fish, but carbon-dioxide content inhibit the
is following a shortened path to the absorption of oxygen. Brockway
outlet screens, obviously the avail- (1950) has shown the same effect
able oxygen will not be utilized effi- for ammonia. The oxygen content
ciently. This condition appears to at the outlet of a pond may not necbe particularly true in both of the essarily be a measure of the availrecirculating types of ponds tested. able oxygen in view of the inhibitShort circuiting appeared to be at ing effect of metabolic wastes on
its minimum in the raceway type absorption of oxygen. An even
of pool under the conditions of the distribution of the fish may protest. It should be emphasized, how- vide a better opportunity for the
ever, that a reduced flow through dissipation of the metabolic prodthe raceway aggravates short cir- ucts, particularly carbon dioxide,
cuiting because of the dead areas and consequently result in a more
that develop along the sides and efficient utilization of oxygen. The
bottom of the pond. The raceway extent and position of eddies in a
and circular pools had almost pond affect not only the disposition
21
of the fish, but also the accumulation of metabolic waste and the
utilization of oxygen.
Carbon dioxide may be present in
the water supply of a pond, and in
large quantities assumes significance. Such a condition, however,
is rare in most water supplies except in instances of multiple reuse
of water without adequate aeration.
Under these circumstances, the picture again is confused by the presence of metabolic products other
than carbon dioxide. Po w ers
(1938) attributes mortalities in
trout to the inability of the fish to
accommodate themselves to abrupt
and repeated changes in carbon-dioxide tension. Marked differences
in carbon-dioxide tensions might be
present in any of the types of ponds
tested, particularly if extensive
eddies are created, as in the FosterLucas pond. The raceway pond,
which is best adapted to the reuse
of water, could produce marked
changes in carbon-dioxide tension
between surface and bottom areas
in the specific instance of inadequate water inflow. The circular
pool, with its sweeping current
along the bottom, would be the least
subject to marked variations in
carbon-dioxide tension. Whether
Powers' premise that the inability
of fish to accommodate to changes in
carbon-dioxide tension is the cause
of mortality, or the Bohr effect (the
inhibition of oxygen absorption) is
the cause, the fact remains that high
carbon-dioxide concentrations can
be detrimental to the well-being of
fish.
The accumulation of metabolic
waste products can produce an un22
favorable environment in rearing
ponds which does not necessarily result in the death of the fish, but is
indicated by their reduced growth
rate or lowered resistance to disease. Brockway (1950) described
the effects of accumulations of
metabolic products measured by
ammonia concentrations. Here
again, reuse of water and inadequate inflows caused unfavorable
conditions for the fish. Although
not specifically stated, Brockway's
description of conditions indicates
that rac,eways were the type of pond
in which he observed high ammonia
concentrations.
The reduction in water level that
Brockway recommends is not necessarily designed to increase the rate
of interchange, but to prevent formation of dead areas along the sides
and bottom of the pond. In the
circular-type pond, the sweeping
action of the current along the bottom of the pond prevents high concentrations of metabolic products.
Even the torus eddy is not susceptible to a buildup of waste products,
because as the products of metabolism settle in the eddy they are
picked up by the bottom current
and carried to the screens. The
current pattern, therefore, may be
an effective substitute for water inflow in some types of ponds.
In the Foster-Lucas pond, location of the eddies is such that movement of the water out of the eddies
is directed, not toward the screens,
but into the main current of the
• pond. This type of current pattern
is conducive to the accumulation
rather than the dissipation of metabolic waste products.
It may be concluded from the
preceding discussion that the carrying capacities of the three types of
ponds cannot be compared without
certain qualifications. Under optimum operating conditions with an
adequate water supply, the hydraulic and biological determinations indicated that the raceway is superior
to the circular pond and it, in turn,
is superior to the Foster-Lucas
pond. With a limited water supply utilized at maximum efficiency,
the circular pool is superior to
either the raceway or Foster-Lucas
ponds. The reuse of water does not
appear advisable, but when such a
procedure is necessary the circular
pool would be superior to the raceway type if the installation could be
such as to secure adequate head between ponds. The Foster-Lucas
pond is not adaptable to the reuse
of water.
DISEASE INHIBITION
It has been demonstrated in actual production operations that under comparable conditions certain
types of ponds are more resistant
to disease development than are
others. In experimental tests, the
raceway exhibited a greater inhibitory effect on disease than did
the circular pool, and the circular
pool a greater inhibitory effect than
the Foster-Lucas pond. These tests
were conducted at the Leavenworth
Station (Wash.), where bacterial
gill disease appears to be endemic
in the water supply and routine
weekly prophylaxis is necessary to
prevent the disease from reaching
epidemic proportions in FosterLucas ponds. Under normal oper-
ating conditions without weekly
prophylaxis at temperatures approximating 60° F. and with fish
loads approximating one-third
pound per cubic foot, blueback
salmon (Oncorh,yneb,us nerka) contracted gill disease in epidemic proportions in 2 weeks' time. Under
comparable conditions, but with
stockings of approximately 1 pound
of fish per cubic foot, 3 weeks were
required for an epidemic to develop
in an 18-foot circular pool. In an
improvised raceway and in deep
troughs with sufficient water inflow
to prevent stratification and poundages not in excess of 1 pound of fish
per cubic foot, the disease did not
reach epidemic proportions and
treatments for the control of the
disease were not necessary. If
water inflows had been reduced in
the raceway type proportional to
those in the circular or Foster-Lucas
ponds—which inflows were comparable—stratification would have
developed in the raceway and the
disease-inhibition characteristic of
this type would have been altered.
The principal factors responsible
for disease inhibition appear to be
isolation of the infected fish and
rapid elimination of disease organisms from the pond. Flow pattern
and the extent and location of dead
areas influence the disease-resistance characteristics of a pond. In
recirculating ponds of the circular
and Foster-Lucas types, there is
little opportunity for the infected
fish to isolate themselves from the
healthy stock. Active fish move
freely throughout these ponds and
pass through the eddies in which
the sick fish concentrate. In the
23
raceway type, healthy fish concentrate principally in the area of inflowing water, while infected stock
tend to collect near the outlet. As
the disease organisms are usually
waterborne, the possibility of infection of healthy stock is materially reduced in the raceway pond
because of the isolation of the sick
fish—assuming, of course, that the
water is not reused and that the inflow is sufficient to prevent stratification.
The extent and location of eddies
are of particular importance in the
evaluation of the disease-inhibition
characteristics of a pond type. Infected fish and disease organisms
tend to concentrate in these eddies.
If the flow pattern is such that
water from the eddy moves directly
to the outlet, the chances of the infection reaching epidemic proportions are materially reduced. Under optimum conditions of water inflow, the raceway pond has the
smallest amount of dead area of any
of the three types tested and this
area is located at the bottom and toward the lower end of the pond. In
the circular pool, the large torusshaped eddy does not extend to the
bottom of the pond, and the sweeping action of the bottom current toward the screen tends to discourage
the accumulation of waterborne
disease organisms. The eddy in the
circular pond is also a true eddy in
that the water is in rapid motion,
which is unfavorable for concentration of sluggish fish. The FosterLucas pond has the largest dead
area, located adjacent to the center
wall with the interchange from this
eddy feeding into the main path of
24
water flow and over the normal concentration points of healthy fish.
This eddy has an imperceptible
current and is a collection point for
infected fish and a concentration
point for disease organisms.
The hydraulic characteristics of
a pond and its disease-inhibition
characteristics appear to be closely
correlated. Alterations in present
pond designs or development of
new pond types may be evaluated
with respect to disease resistance by
determination of the hydraulic
characteristics of the structure in
model studies.
FOOD DISTRIBUTION
Food distribution in a pond is
primarily a function of current velocity. In general, recirculating
types of ponds develop a higher
current velocity than raceway
ponds, and are more efficient in food
distribution. Palmer et al. (1951) ,
in experiments on frequency of
feeding, concluded that concentrations of fish and food are the principal factors responsible for food
wastage when large fingerlings are
fed. Ponds that have the fish concentrated in one particular area and
that do not have sufficient current to
distribute the feed rapidly should
be fed at more frequent intervals
and more slowly than ponds that
have the fish more evenly distributed and that have sufficient water
currents to distribute the feed more
rapidly, if food wastage I; to be
reduced to a minimum.
A current of 0.8 to 1 foot per
second will distribute most sinking
feeds throughout the periphery of
a circular pool and throughout the
narrow section of a Foster-Lucas
pond. Both of these ponds had currents of this magnitude. The raceway had a current of less than 0.1
foot per second. In such a current,
sinking feeds drop to the bottom
immediately, and food distribution
must be accomplished artificially by
introducing the food along the
length of the pond. This procedure
is more timeconsuming and, therefore, less efficient.
Floating foods present a slightly
different problem in that it is desirable to keep them from coming in
direct contact with the turbulence
of the inflow to prevent their excessive leaching. In the circular pool,
floating foods are introduced at a
single point at the periphery of the
pond where the food will make almost a complete revolution before
it is subjected to the action of the
inflow. About 8 minutes are required for a single revolution and
during this period the bulk of the
food is consumed. Two points of
introduction are used in the FosterLucas pond, downstream from each
header pipe. In this manner, the
food is consumed before it passes
under the opposite inflow pipe. In
the raceway, with its inflow spill,
the food should be introduced below the point of excessive turbulence and for at least 20 feet downstream to ensure adequate distribution. Because of the slow current,
there is little distribution even of
floating foods in the raceway type
of pond.
The more effective the food distribution by the pond the less time
is required for feeding, and the
more efficient is the pond. Using
this criterion, the circular pool is
superior to the Foster-Lucas pond,
and the Foster-Lucas in turn is
superior to the raceway.
CLEANING EFFICIENCY
The cleaning efficiency of a pond
is governed by current velocity,
flow pattern, and location of the
eddies. A tremendous difference
in cleaning efficiency was found to
exist between the three ponds
studied.
Current velocities of 0.8 to 1 foot
per second, or more, carry excrement and all but the heaviest debris. As the velocity decreases,
the heavier particles settle out until
at about 0.1 foot per second all but
the most semibuoyant particles are
deposited. In the circular pool,
current velocities were sufficiently
high so that no deposition of excrement or debris occuirred, except at
the screens. The ' Foster - Lucas
pond had a range in current velocity
from 0.8 foot per second to practically zero, and debris was deposited in these areas of low velocity.
In the .raceway, except in the area
of turbulence at the inflow, the current velocity averaged less than 0.1
foot per second. As a result, a continuous settling action occurred
along practically the entire length
of the pond.
If the flow pattern in a pond is
such that the eddies are located adjacent to the outflow or the current
sweeps the material dropped by the
eddies into the outflow, the selfcleaning characterists of the pond
are not sacrificed. In the circular
pool, the latter case was true; the
25
sweeping action of the bottom current picked up the debris as it settled from the torus-shaped eddy
and deposited it at the screens.
The flow pattern of the FosterLucas pond was such that the
debris and excrement were deposited in the large eddy adjacent
to the center wall with the principal
deposition at the maximum distance from the outflow. When the
pond was drawn down during
cleaning, the sediment was picked
up by the main-current flow and recirculated. The most effective
method of cleaning this pond was
to keep the debris stirred up and
rely on the drawdown to remove it.
Such a procedure proved both slow
and inefficient in that it was impossible to remove all the material at
one cleaning. The raceway had little self-cleaning action because of
the low current velocity, but the
material could be brushed toward
the screens as the pond was drawn
down. With large fish it was possible to reduce the water depth for
cleaning and rely on the increased
current and swimming action of
the fish to clean the upper portion
of the pond. Neither the raceway
nor the Foster-Lucas ponds had the
self-cleaning characteristics of the
circular pool, but of the two the
raceway was superior.
COMPARATIVE POND EFFICIENCIES
To determine the efficiency of any
pond type, it is necessary to consider the pond in relation to each
of the four criteria that have been
discussed. In table 2, the relative
standing of the three types, together with their relative efficiencies, has been enumerated. No attempt has been made to weight the
standings, because different water
supplies present different problems.
In regions where endemic diseases
jeopardize operations, the diseaseinhibition characteristic would be
of prime importance in the selection of a pond type. The amount
of water available and the desired
output of the station also influence
the selection. In this instance, an
attempt has been made to rate the
types under conditions of optimum
inflow and comparable inflows.
With optimum inflow, the circular
and raceway types are approximately equal in efficiency, according to this arbitrary rating. If disease problems were anticipated, the
raceway would be selected. Should
Table 2.—Comparative efficiency ratings of three pond types
Carrying capacity
Pond
Raceway
Circular
Foster-Lucas
26
Efficiency
Disease inhibition
Food d is- SelfOptimum Compare,- Optimum Compare- tribution cleaning Optimum Comparewater ble water
water ble water water ble water
supply supply
supply
supply supply supply
1
2
3
2
1
3
1
2
3
2
1
3
3
1
2
2
1
3
1.75
1. 50
2. 50
2.25
1.00
2. 75
the water supply be limited and
maximum output desired, the circular pool would be the type selected.
Several fallacies in this method
of comparative ratings are apparent if the time element is considered.
In recirculating types of ponds, the
time required for feeding and the
frequency of feeding for optimum
food utilization are nearly comparable. In the raceway, both the
time required for feeding and the
frequency of feeding would be at
least twice that in the recirculating
ponds. The time required for
cleaning the circular pool is onethird that for the Foster-Lucas
pond. The time consumed in
cleaning the raceway is slightly less
than in the Foster-Lucas, but much
greater than in the circular pond.
On the other hand, where weekly
or biweekly prophylaxis can be
eliminated, the saving in time
would be considerable. The time
element is, in a sense, an intangible
factor in that the operational policy
of a station determines the time
spent on these various items.
The results of these investigations indicated that none of the
pond types studied even approaches
perfection. Neither can one type
of pond be recommended as superior to the others under all circumstances. One fact is evident : the
Foster-Lucas type of pond is inferior in every instance to either
raceway or circular pools when
measured by the four criteria used
in making these evaluations.
These studies have demonstrated
that the physical and biological
characteristics of new pond designs
or alterations in established types
of ponds may be evaluated from the
hydraulic characteristics of the
pond. Models may expedite the
determination of these hydraul,ic
characteristics.
SUMMARY AND CONCLUSIONS
Hydraulic investigations were
conducted on three types of rearing
ponds. Models were constructed of
the three prototypes to determine
hydraulic similitude. The flow
patterns in the models and prototypes were observed by the use of
floats and dyes. The degree of
short circuiting and mixing, and
the apparent detention time and
flowing-through time were determined by injecting dye in the inflow
and measuring both the time of its
appearance and its concentration at
the outflow.
Flow-pattern determinations in
the models and prototypes dis-
closed the following differences in
character between pond types:
1. In the Foster-Lucas pond
studied, there was a recirculating
current of 0.8 foot per second, a
large eddy behind the partition
wall, a roll at each end of the pond,
and short circuiting.
2. In the circular pond, there was
a recirculating current of 1 foot per
second at the periphery, a large,
peripheral mixing zone, a torusshaped eddy, and an inflowing
bottom current that contributed to
short circuiting.
3. In the raceway pond, there
was no recirculating current, a
27
mean current velocity of less than
0.1 foot per second, turbulence at
the inflow caused by a high inlet
velocity, and a high water demand
to prevent stratification.
Studies in dye-concentration time
produced the following comparative evaluations :
1. The Foster-Lucas and circular
ponds had dead areas comparable
in extent. In the raceway, these
dead areas were much reduced.
2. There was more interchange
between the dead areas and the
main current in the Foster-Lucas
type than in the circular pond.
3. Short circuiting was greatest
in the Foster-Lucas, less in the circular, and least in the raceway type
of pond.
Model studies indicated that the
hydraulic characteristics of fishrearing ponds may be studied satisfactorily in 1: 10 scale models.
Hydraulic conditions in the three
types of ponds were correlated with
carrying capacity, disease inhibition, food distribution, and cleaning efficiency, as follows:
1. Carrying capacity was correlated with the available oxygen,
carbon-dioxide content, and the accumulation of metabolic waste
products.
(a) The amount of available
oxygen was affected by the volume
of water inflow, the flow pattern,
28
short circuiting, and distribution of
the fish.
(b) The carbon-dioxide content
in the ponds not associated with
metabolic waste was correlated with
inadequate aeration and the reuse of
water. Changes in carbon-dioxide
tension in a pond were influenced
by the location and extent of the
dead areas.
(e) The accumulation of metabolic waste was affected primarily
by the extent and location of dead
areas.
2. The disease-inhibition characteristic of the three pond types was
correlated with the degree of isolation of infected stock as influenced
by the flow pattern, and the rapid
elimination of disease organisms as
affected by the extent and location
of the dead areas.
3. Distribution of food in a pond
was indicated to be primarily a
function of current velocity.
4. The great variation in cleaning efficiency of the three ponds
studied was attributed to differences in current velocities, flow patterns, and location of the eddies.
Of the pond types studied, no
single type could be considered superior under all conditions of fishcultural operations. The FosterLucas type, however, was definitely
inferior to either the raceway or
circular pond.
LITERATURE CITED
C., and VIRGINIA S. BLACK.
1950. Carbon dioxide asphyxiation of some British Columbia fresh-water fishes.
American Jour. Physiology, vol. 163, No. 3, p. 699. December. [Abstract.]
BROCKWAY, DONALD R.
1950. Metabolic products and their effects. U. S. Department of the Interior,
Fish and Wildlife Service, Progressive Fish-Culturist, vol. 12, No. 3, pp. 127-129.
July.
COBB, DIEN W., and J. W. TITCOMB.
1930. A circular pond with central outlet for rearing fry and fingerlings of the
Salmonidae. Trans. American Fisheries Society, vol. 60, pp. 121-123.
BLACK, EDGAR
DAVIS, HERBERT SPENCER.
1946. Care and diseases of trout. U. S. Department of the Interior, Fish and
Wildlife Service, Research Rept. 12, 98 pp., illus., 3d revision.
ELLIS, M. M., B. A. WESTFALI, and MARION D. ELLIS.
1946. Determination of water quality. U. S. Department of the Interior, Fish
and Wildlife Service, Research Rept. 9, 122 pp.
IRVING, L., E. C. BLACK, and V. SAFFORD.
1941. The influence of temperature upon the combination of oxygen with the
blood of trout. Marine Biological Lab., Biol. Bull., vol. 80, No. 1, pp. 1-17.
February.
JOHNSON, HARLAN E., and ALFRED C. GASTINEAU.
1952. A comparison of the growth of fingerling chinook salmon reared in ponds,
troughs, and circular tanks. U. S. Department of the Interior, Fish and Wildlife Service, Progressive Fish-Culturist, vol. 14, No. 2, pp. 76-78. April.
PALMER, DAVID D., H. WILLIAM NEWMAN, ROBERT L. AZEVEDO, and ROGER E. BuRRows.
1952. Comparison of the growth rates of chinook salmon fingerlings reared in
circular tanks and Foster-Lucas ponds. U. S. Department of the Interior, Fish
and Wildlife Service, Progressive Fish-Culturist, vol. 14, No. 3, pp. 122-124.
July.
PALMER, DAVID D., LESLIE A. ROBINSON, and ROGER E. BURROWS.
1951. Feeding frequency : its role in the rearing of blueback salmon fingerlings
in troughs. U. S. Department of the Interior, Fish and Wildlife Service, Progressive Fish-Culturist, vol. 13, No. 4, pp. 205-212. October.
POWERS, EDWIN B.
1938. Factors involved in the sudden mortality of fishes. Trans. American
Fisheries Society, vol. 67, pp. 271-281. 1937.
PHEVOST, GUSTAVE.
1941. A method of increasing the capacity of a trout hatchery. Trans. American Fisheries Society, vol. 70, pp. 430-435. 1940.
RODGERS, ERNEST 0.
1949. A drop center raceway. U. S. Department of the Interior, Fish and Wildlife Service, Progressive Fish-Culturist, vol. 11, No. 3, pp. 189-190. July.
SURBER, EUGENE W.
1936. Circular rearing pools for trout and bass. U. S. Department of Commerce,
Bur. Fisheries, Progressive Fish-Culturist, No. 21, pp. 1-14.
29
U.S. GOVERNMENT PRINTING OFFICE 1955 0-334317